TranscriboQuest 2025: Arabic Team - Segmentation data

The dataset produced during the task for the TranscriboQuest Arabic team 2025. The task focussed on creating segmentation data for Arabic manuscripts with complex layouts. The repository contains 3 file types: TranscriboQuest_2025_Arabic_metadata.csv - A csv file documenting the manuscripts with their call numbers, catalogue links and file names for the data. Transcriboquest-Arabic-documentation.md - documentation detailing the task and the guidelines that were followed. Zip files for each manuscript containing ALTO xmls and page images. Note: This dataset is a segmentation dataset (lines and zones). Not all xmls have corrected transcriptions.